Introduction

II-Agent updates and hosted web app beta are now live. These updates take the original II-Agent prototype you’ve seen and refines it into a fast, reliable product for everyday use.

Waitlisted users can now try the web app on agent.ii.inc.

Not on the waitlist? You can self-host the new II-Agent today, see public repo.

Why a Web App, and Why Now

The initial II-Agent release proved that agentic workflows can dramatically accelerate technical work, but our community told us they needed less setup, more stability, and a unified place to run complex tasks end to end.

The new web application with a fresh UI delivers exactly that: instant access, consistent environments, and a design that surfaces the most important actions without friction.

If you loved the original vision, you’ll find even more power here, now simplified, hardened, and ready for production-grade use.

What’s New in This Release

Streamlined access and model freedom

With the web app beta, there’s nothing to install, and you can Bring Your Own Key (BYOK) to try any model you want. Choose the models that fit your budget, latency needs, or quality bar. Your keys, your call.

Purpose-built workflows

We’ve focused the experience around three high-impact pillars: full-stack website creation, deep research, and slide creation. That means fewer menus and more momentum. Start a session, pick your goal, and II-Agent assembles the right tools, prompts, and checks to get you from idea to output fast.

Safe extensibility with your tools

Every team brings its own stack. In this release, you can plug in any MCPs your workflow requires, all backed by a standalone E2B sandbox for safe, isolated execution. This keeps experiments contained, secrets protected, and results reproducible, whether you’re testing a scraper, building a data pipeline, or evaluating a custom retrieval tool.

Collaboration with Codex Agent

For complex, code-heavy tasks, II-Agent now collaborates with Codex Agent, combining II-Agent’s orchestration and workflow memory with Codex’s autonomous coding strengths. You can hand off targeted subtasks (for example, refactors, API integrations, unit test generation) while II-Agent manages the broader plan, checkpoints, and artifact assembly.

The result is deeper automation where it matters, with a clean review loop and the option to BYOK so you can mix and match models and Codex-driven steps in one seamless run.

How It Helps: Four High-Impact Scenarios

1. Full-stack website creation

From blank brief to deployed app, II-Agent scaffolds frameworks, wires components, and iterates on UX with reviewer feedback. Generate a Next.js or React front end, connect a FastAPI back end, and set up a database in minutes, not days. Because execution happens in an isolated E2B sandbox, you can safely test integrations and version changes before shipping. When you’re ready, promote to your preferred environment.

2. Deep research that stays grounded

Complex questions require more than a single search. II-Agent plans a research path, pulls and ranks sources, extracts claims, and compiles a structured report, complete with citations and notes for follow-up. You choose the model via BYOK, mix speed and accuracy as needed, and keep the full trace of what was searched and why.

3. Slide creation that doesn’t derail your day

Turn structured content (findings, metrics, product specs) into clean, on-brand decks. The agent proposes layouts, drafts speaker notes, and iterates on tone and emphasis. If you need custom data pulls or charts, attach an MCP that fetches them, run it in the sandbox, and drop the results straight into slides.

4. Code work with Codex Agent in the loop

When you need heavy lifting on code, bring Codex Agent into the conversation. II-Agent maintains the plan, sets constraints, and validates outputs. Codex Agent focuses on implementation details, writing functions, fixing failing tests, or instrumenting telemetry. You retain oversight with readable logs, diff views, and retry controls.

Try It Today

These updates transform our prototype into a daily driver, offering faster builds, deeper insights, and smoother storytelling, now with Codex Agent as a first-class collaborator.

If you were on the waitlist, go to agent.ii.inc to try the beta web application and see details on our website.

Otherwise, head over to GitHub and run it yourself.

If you’re coming from the original project, start here to see how far the experience has come.